Handwriting_recognisation

Description :

• Handwriting recognition for Cursive and Non-cursive using Tesseract-OCR .
• Image Processing techniques(method of histograms) for cursive segmentation .
• Spline curves for smooth data points interpolation.
• Simulator made from Tkinter GUI in Python .
• Processing and OpenCV code in C++ which is executed through Python wrapper

User Manual :

• Step 1) Install tesseract , PIL and OpenCV in Ubuntu (follow SETUP_LINKS.docx)

• Step 2) To run GUI :

  • python tkinter_image.py

  • i . Draw the image
  • ii. Click "Save" to save the image and coordinates :: - a. Image saved as "Image.jpg" - b. Coordinates saved in "letter_sampleNew.txt"
  • iii. Click "Clear" to clear the image (Optional)

• Step 3) To refactor the coordinates and display it :

  • i. Click "Factor $ Save" - a. Factored coordinates saved in "letter_factor.txt

• Step 4) // To be run only for cursive writing

  • To run the OpenCV- segmentation code
    • i. Click "Segment_OpenCV"

The final segmented image is saved as "ImageFinal.jpg"

• Step 5) To show the recognisation using Tesseract

  • i. Click "Cursive" for recognisation of cursive word(after step4)

  • ii. Click "Character" for recognisation of single charact(ignore4)

  • iii. For non-cursive word :

    - a. In new terminal goto required directory :
    
     > tesseract Image.jpg out -l eng -psm 6
    
    - b. Open "out.txt" , ignore step 6 
    

• Step 6) Open Output terminal to get the recognised word or Open file "tess_output.txt" to get the recognised word

• Step 7) Click "Clear" to remove everything and redraw

Simulator :

Simulator :

Segmented word from OpenCV :